This interview covers how early-stage investor Elad Gil spots promising markets. He believes market matters more than team—a great team in a bad market will lose. He's bullish on machine learning chips, anti-aging, and turning corporate tasks (like background checks) into API services. Key holdings: PagerDuty (early on, no sales team, but Amazon and Apple were customers), Snapchat (rough UI but users kept using it), Uber (a few heavy users could cover a city's costs).
Elad Gil shared his investment methodology for identifying interesting markets on the Invest Like the Best podcast. The core thesis is that evaluating early-stage companies should focus on product-market fit rather than relying solely on data metrics. He recommends using reference checks to verify t
Elad Gil is a serial entrepreneur and early-stage investor with a PhD in biology. He has participated in investments in Airbnb, Coinbase, Stripe, and others, and helped Twitter scale from 100 to 1,500 employees. This issue focuses on how to identify promising early-stage markets and companies. Gil argues that in early-stage investing, market matters more than team—"In a bad market, even the best team will lose; in a great market, even the worst team can win" —a judgment that directly challenges the mainstream consensus of "team first" in the early-stage investment community.
Gil believes that early-stage investment evaluation should be divided into three layers: market first, team second, and personal preference third.
Data-driven success signals: Reviewing data from his own portfolio companies, Gil summarizes four early signals:
1. Product already built (even if terrible) — Companies with a demo rather than just a PPT perform better
2. Organic growth — Even with a small base, 20-30% monthly growth via word-of-mouth is a strong signal
3. Enterprise customers using it spontaneously — For example, PagerDuty had no sales team early on, only 4 engineers, yet Amazon and Apple were already its customers
4. Product is terrible but people still use it — For instance, early Snapchat had a rough UI, but users kept using it
> Quote: "If you have a great team in a terrible market, the market wins. If you have a terrible team in a great market, the market wins. And if you have a great team in a great market, something magical happens." — Meaning: a great team in a terrible market, the market wins; a terrible team in a great market, the market wins; a great team in a great market, magic happens.
Gil argues that most tech companies fail not because of a bad product, but because they underestimate the importance of distribution.
1. Distribution – Continuously broaden and saturate distribution channels (e.g., Google paid to acquire clients through the Firefox toolbar, Adobe bundling, etc.).
2. M&A – Entering new areas through acquisitions.
3. Internal Product Iteration – Continuously launching new product lines.
> Quote: "Companies that tend to sort of under-execute or end up with a $5 billion company, which is amazing, but you could have been a $50 billion company. Often it's because they weren't aggressive enough on three things."
Gil argues that using average customer lifetime value (LTV) versus customer acquisition cost (CAC) to evaluate a business can miss critical signals, and that the economics of top-tier "whale users" should be the focus.
Gil proposes three types of "non-obvious" markets and identifies three currently undervalued investment directions.
Three Market Types:
1. Entirely new markets with extremely rapid growth — the rarest, e.g., cryptocurrencies, internet browsers in the 1990s
2. Seemingly crowded but actually early-stage — e.g., when Google started, there were already a dozen search engines; when Dropbox and Box started, there were already a dozen cloud storage services. These pioneers identified real demand, but their products failed to meet it, leaving room for later entrants
3. Breakthroughs in distribution or channels — e.g., Zynga rose by leveraging Facebook's new distribution channel; SaaS companies entered from the low end and moved upmarket; Tesla entered from the high end and moved downmarket. Analogy: the "mini-mill" in the steel industry started with scrap metal recycling and eventually became the market mainstream
Three Currently Undervalued Markets:
1. Machine learning chip layer — building custom ASICs to replace NVIDIA GPUs (GPUs are not optimized for ML). Gil believes this market "can produce at least one company worth tens of billions of dollars." Analogy: every technology wave spawns a major chip company — Qualcomm/ARM for the mobile wave, Broadcom for the networking wave, Intel/AMD for the PC wave
2. Longevity/anti-aging field — based on Gil's biology background, he believes there are 20-30 years of genetic and pathway data (e.g., mTOR pathway, rapamycin, metformin, etc.) that have not yet been fully commercialized
3. "Deconstructing Fortune 500 companies" opportunity — turning repetitive tasks within large companies (e.g., background checks) into independent API services. For example, Checkr upgraded background checks into an API service, first serving the gig economy and then expanding to other areas
Gil argues that the two most overlooked issues for CEOs are "burnout from doing things they hate" and "inability to let go/delegate."
1. Spending a significant amount of time doing things they hate (e.g., a product-focused CEO getting bogged down in sales compensation discussions or HR matters)
2. Inability to let go or feeling challenged by newly hired executives
Perspective on organizational structure: Gil is skeptical of "decentralized" organizations (such as Holacracy or the Valve model). He believes humans have not fundamentally changed in 10,000 years—they need direction, coordination, and processes. "If you're making dinner with a few friends, you still have to decide who brings the salad, who makes the main course, and who brings dessert; otherwise, you'll end up with five desserts." He adds that some companies succeed despite poor organizational structures because their product-market fit is so strong—"they succeed despite themselves, not because of it." If properly organized, these companies could have been ten times more successful.
| Position | Guest Stance | Key Data |
|---|---|---|
| PagerDuty | Bullish (early-stage investment case) | Initially only 4 engineers, no sales team; Amazon and Apple were already customers |
| Snapchat | Bullish (early signal case) | Rough UI but users continued to use it |
| Uber | Neutral (anecdotal reference) | In early days, 5-10 "whale users" in a single city could cover office operating costs |
| BioAge | Bullish (personal involvement) | Andreessen Horowitz led the Series A |
| Spring Discovery | Bullish (personal involvement) | Applied machine learning to longevity testing |
| Checkr | Bullish (market opportunity case) | Background check API company, initially serving the gig economy |
| Bitmain | Neutral (industry analogy) | Cryptocurrency ASIC chip company |
| Unity Biotechnology | Neutral (industry mention) | Focused on senolytics (anti-aging drugs), may have gone public |
| Novartis | Neutral (industry mention) | Conducted immune response trials in the elderly using rapamycin |
1. "Market First, Team Second" — Practitioner of Rachleff's Law: Gil argues that in early-stage investing, the market matters more than the team, contradicting the consensus of most angel investors who prioritize "team above all." Support: Reviewing his own portfolio, he found that the commonality among successful cases was genuine market demand, not a glamorous team background.
2. Negative background checks are a neutral signal, not a negative one: People can behave entirely differently in different environments. Support: A colleague deemed "lazy" on Twitter later founded one of the most successful startups in that community, because he "finally felt his own ass was on the line."
3. "Product is king" is a disaster; distribution is the real bottleneck for growth: Most companies have only one effective distribution channel. The key to scaling from $5 billion to $50 billion lies in the aggressiveness of three things: distribution, M&A, and product iteration. Support: Google relied on word-of-mouth early on but quickly shifted to partnerships and paid client distribution.
4. Evaluating a business with average LTV/CAC misses key signals — focus on "whale users": The top 5-10 users may cover the operating costs of an entire new market. Support: An early rumor about Uber — a handful of high-frequency black car users could sustain a city office.
5. Three non-obvious market frameworks: Entirely new high-growth markets, seemingly crowded but actually early-stage (products not meeting demand), and distribution/channel breakthroughs. Support: Google started when there were already a dozen search engines; mini-mills began with scrap steel recycling and became mainstream.
6. Every technology wave spawns a major chip company: Mobile → Qualcomm/ARM, Networking → Broadcom, PC → Intel/AMD. The ML chip layer "can at least give birth to a $10 billion company." Support: GPUs are not optimized for ML; custom ASICs will significantly outperform in power consumption and performance.
7. "Decentralized" organizations (Holacracy/Valve model) are fundamentally flawed: Humans need direction, coordination, and processes. Support: Some companies succeed despite terrible organizational structures because product-market fit is too strong — "they succeed despite themselves; if organized properly, they could have succeeded 10 times over."
8. The biggest hidden killer for a CEO is doing things they hate: A product-focused CEO getting bogged down in sales compensation discussions or HR matters can lead to burnout; inability to let go/delegate is the second biggest breaking point for first-time founders. Support: Gil has observed the growth trajectories of numerous founders.